This invention relates to photolithographic compositions employed in the manufacture of microelectronic devices, their method of manufacture, and particularly to methods for applying these compositions as multifunctional thin films onto semiconductor substrates.
In the manufacture of micro electronic devices, photolithographic printing is employed to fabricate circuit pattern images onto semi-conductor substrates. In this process, photosensitive films called photoresist are coated onto the substrate, exposed to light, and then developed in an alkaline developer solution. Upon development, a pattern configuration forms in the photoresist corresponding to a change in solubility of those regions of the photoresist material exposed to the irradiating light. The clarity or resolution of the lines which define these patterns at microns or even sub-micron geometries to a great extent serves as a limitation to the photolithographic process. However, photolithographic technology is approaching its ultimate limit, the point beyond which resolution cannot be improved due to diffraction effects, incompatibility of materials, and complexity of processing.
For example, one of the problems which exists in processes of this type is called reflectivity. This is caused by the fact that some of the light striking a thin layer of photoresist material will usually pass through the layer and be reflected upward from the substrate during the radiation exposure. As the incident light is generally not perfectly normal to the surface of the topography of the substrate, and as it may be diffracted upon passage through the photoresist, the incident light will be reflected angularly from the surface of the substrate rather than normally therefrom. Such light will impinge upon the unexposed portions of the photoresist and some may again pass through the photoresist to strike the opaque portions of, for example, a photo mask, and this light will be reflected back into some portions of the photoresist which are not intended to be exposed. As a result of light being reflected back and forth between the substrate and the photoresist, as well as light being scattered from surface irregularities, there may be a pronounced detrimental effect upon the ultimate resolution which can be obtained upon photo development. These problems are even more pronounced by the standing wave phenomenon and/or the reflective notching phenomenon experienced when patterning or photo developing the material.
Previous attempts to correct the reflectivity problems as disclosed in, for example, U.S. Pat. No. 4,102,683, call for interposing a light absorbing layer between the surface of the substrate and the photoresist material. These so-called anti-reflective layers have the property of absorbing light which passes through the photoresist and not reflecting it back upward. Such a light absorbing layer may be manufactured by mixing a fluorescent dye such as an organic phosphor with an organic binder such as Fluorel made by the 3-M Company or Viton made by DuPont which is a rubber.
Another prior art technique employed to correct reflectivity problems involves the use of a quarter-wave plate having an odd multiple thickness of one-quarter of the wavelength of light to which the photoresist layer is sensitive. This plate is comprised of silicon dioxide thermally grown or deposited in any manner such as by the decomposition of silane. When two light waves of the same amplitude and wavelength arrive at a given point simultaneously and in phase, illumination will occur at that point. If they are out of phase, no illumination will occur because the waves annul each other. By controlling the thickness of the quarter wave plate one can throw reflected light out of phase with incoming light and cancel the reflected light from the substrate.
Other prior art describing similar methods for addressing this reflective phenomenon have been discussed: M. Listvan, et al., in their text "Multiple Layer Techniques in Optical Lithography: Applications to Fine Line MOS Production", published in S.P.I.E., Volume 470, 1984, p. 85, and by R. Coyne, et al., in their article "Resist Processes on Highly Reflective Surfaces Using Anti-Reflective Coatings", published in the proceedings of the Kodak Micro-Electronics Seminar, Interface 1983. Additionally, K. Polasko, et al., discusses this phenomenon in their article "Thin Silicon Films Used as Anti-Reflective Coatings for Metal Coated Substrates", published in S.P.I.E., Volume 631, 1986, p. 181. However, such prior art anti-reflective layers have exhibited a number of problems not the least of which is the fact that in general, when the surface of the substrate is irregular, there is required an additional planarization material or layer. Other disadvantages are, for example, that when prior art organic binders have been patterned by a wet etch development, such layers develop isotropically resulting in undercutting during development which results in a narrow processing latitude or even complete lift-off of sub-micron geometries. Also, certain organic binders such as PMMA and polybutene sulfone when employed as an anti-reflective sub-layer have poor stability as vehicles for etching patterns onto the substrate.
As previously mentioned, when the surface topography of the substrate is irregular, a planarization layer has sometimes been employed which may or may not contain anti-reflective material. Such planarization layers reduce the significance of depth of focus limitations of patterning equipment. In particular, an imaging surface must be flat because the optical patterning equipment can focus only in one plane and will not readily refocus microimages to adjust for nonflatness of a substrate. The depth of focus of an optical printer then becomes a limiting parameter. Since depth of focus is a function of the thickness variation, when using thinner resist or attempting to achieve higher resolution, small thickness variations become more significant and the depth of focus is even more limiting. By planarization consistent thickness and minimal variation can be attained, thereby eliminating the depth of focus problem.
Materials employed as planarization agents are, for example, polymethyl methacrylate (PMMA), polyimides, or phenolformaldehyde condensation resins such as Novolak. These planarization layers, however, in turn require the use of additional layer materials such as adhesion promoters to assist in the layer adhering to the substrate, and for example, interfacial barrier layers between the planarization layer and the photoresist absent a bake prior to overcoating with a photoresist. Such planarization layers have been disclosed in, for example, U.S. Pat. No. 4,370,405, and U.S. Pat. No. 4,524,121.
Interfacial mixing of the photoresist layer, and prior art sub-layer materials, particularly PMMA, is detrimental to the ultimate resolution desired. This stems from the fact that prior art sub-layer polymers cannot withstand overcoating of the photoresist without degrading the film's integrity. Accordingly, interfacial layers have been disclosed comprised of, for example, polyvinyl alcohol polymers, and polyimide precursors. Such interfacial barrier layers have been disclosed in H. Ohtsuka, et al., in their article "PCM Resist Process With RIE Development Method Applied for the Aluminum Etching Process", S.P.I.E., Volume 631, p. 337, 1986; and disclosed in C. Ting, et al., in their article "An Improved Deep Ultra-Violet Multi-Layer Resist Process for High Resolution Lithography", S.P.I.E., Volume 469, p. 24, 1984.
Additionally, it has been suggested in the prior art to employ a lift-off or release layer which is composed of, for example, thick films in the 1 to 3 micron range made from polysulfone polymers, polyimides specially fabricated, or other extraneous photoresist materials. This layer is applied in multi-layer lithography. After pattern transfer, such release layers can be dissolved or physically expanded in its solvent to cause release of all layers coated above it.
Examples of this technology can be found in U.S. Pat. No. 4,692,205 and U.S. Pat. No. 3,873,361. These lift-off layers are also taught to require the assistance of an adhesion promoter for both the release layer and the top photoresist material, and an oxygen etch barrier layer between the photoresist and the lift-off layer. Even with the release layer such as described in U.S. Pat. No. 4,692,205, at least two hours have been required as a practical matter for the lift-off of the layers in this system.
To improve the line resolution, given this technology, there has also been proposed surface application of a contrast enhancement layer onto the photoresist. These contrast enhancement layers have been disclosed by B. Griffing, et al., "Application of Contrast-Enhanced Lithography to 1:1 Projection Printing", S.P.I.E., Volume 469, p. 94, 1984; K. Patrillo, et al., "CEL Resist Processing for Sub Micron CMO's and Bi-Polar Circuits", S.P.I.E., Volume 920, p. 82, 1988. They disclose that these enhancement layers are photobleachable dyes in an inert resin and absorb the light diffracted from the edge of openings in a photo mask used to pattern the photoresist. The enhancement layer increases contrast, which, in turn, increases sidewall angle and the minimum resolution capable from a particular exposure system.
The adhesion promoters previously mentioned in passing are required to provide the best possible adhesion between the substrate, the photoresist layer, and other layers employed in this technology. When conducting photolithography to manufacture micron and sub-micron patterns, adhesion between the photoresist layer and the substrate must be maximized. The extremely minute area between the photoresist and the substrate and the harsh processing conditions subsequent to the application of the photoresist, render adhesion a critical parameter of the process. However, adhesion promoters are usually silicon based materials applied either by spin coating or vapor application as disclosed in U.S. Pat. No. 3,549,368 and U.S. Pat. No. 3,586,554.
The prior art has taught the application of such adhesion promoting materials in a molecular mono-layer and are limited as to what substrates they will compatibly affect adhesion, varying, for example, as between silicon oxide, silicon itself, or silicon nitride. Effective adhesion by an adhesion promoting layer thicker than 100 angstroms has not been possible.
Although the above-described special layers have provided multi-layer material which solves a number of problems, the industry has been dissuaded from perfecting photolithographic processes at sub-micron levels because of the limitations with regard to these materials and the difficulty in processing these multi-laminates.
For example, not only does multi-laminate application require tedious and precise control of the individual film thicknesses, but moreover, many of the layers, particularly those previously used as anti-reflective layers, planarization layers, release layers, etc., require individual baking prior to the application of other layers which is time consuming and oftentimes requires additional equipment.
Accordingly, a new and improved composition which could be applied as a single multifunctional layer providing anti-reflectivity, adhesion, releasability, contrast enhancement, and yet could inhibit interfacial degradation of the sub-layer integrity without the necessity of thermally baking or curing the composition prior to applying a photoresist material would be a welcomed and unexpected advancement in the art.